From owner-freebsd-questions Tue Oct 29 1:58:52 2002 Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 9326537B401 for ; Tue, 29 Oct 2002 01:58:50 -0800 (PST) Received: from aueb.gr (hermes.aueb.gr [195.251.255.142]) by mx1.FreeBSD.org (Postfix) with ESMTP id 92D3243E3B for ; Tue, 29 Oct 2002 01:58:47 -0800 (PST) (envelope-from dds@aueb.gr) Received: from aueb.gr (spinellhw-1.eyelpidwn.offices.aueb.gr [195.251.233.14]) by aueb.gr (8.8.5/8.8.5) with ESMTP id LAA03396 for ; Tue, 29 Oct 2002 11:54:53 +0200 (EET) Message-ID: <3DBE5B60.7F3A7F52@aueb.gr> Date: Tue, 29 Oct 2002 12:56:48 +0300 From: Diomidis Spinellis X-Mailer: Mozilla 4.78 [en] (Windows NT 5.0; U) X-Accept-Language: en,el,de MIME-Version: 1.0 To: freebsd-questions@FreeBSD.ORG Subject: Process inexplicably blocks after upgrading to 4.7-RELEASE Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG After upgrading a diskless client and the respective NFS server from FreeBSD 4.1 to FreeBSD 4.6 (and now 4.7) I have been having trouble with a process on a client blocking for about 15 seconds every 10-20 minutes. The process is an MP3 player (mpg123), so I can not really afford to have it block for that long. The mpg123 process runs on the diskless client, reading data from the server and playing the music locally. Every so often mpg123 stops playing music for about 15 seconds. An strace(1) on the mpg123 process shows it waiting for a write(2) on /dev/dsp to return. If however I break into the kernel debugger on the client and trace the mpg123 stack frame I always find it blocked in the following sequence: sbwait(...) soreceive(...) nfs_send(...) nfs_reply(...) nfs_request(...) nfs_readrpc(...) nfs_getpages(...) vnode_pager_freepage(...) vm_fault(...) trap(...) trap(...) (at that point the stack trace stops due to a "page fault while in kernel mode"). The problem first manifested itself when the client and server were upgraded from FreeBSD 4.1-RELEASE to 4.6.2-RELEASE-p2. The problem also occurs when the client runs 4.7-RELEASE-p1. On the client side the network and ethernet card are configured as follows: ed0 at port 0x240-0x25f iomem 0xd8000 irq 9 on isa0 ed0: address 00:00:e8:32:81:e8, type NE2000 (16 bit) pcm0: at port 0x530-0x537,0x538-0x539,0xf8c-0xf94,0xe0e irq 10 drq 1 flags 0xa110 on isa0 No messages are logged on the client side; on the server I found a few "nfsd send error 64" (EHOSTDOWN) messages, but I think these were generated after staying too long in the kernel debugger. I would really appreciate any ideas on how to resolve this problem, or suggestions on how to further attack it. Thanks, Diomidis - http://www.dmst.aueb.gr/dds/ mailto:dds@aueb.gr To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message